FunASR v1.3.14: safer SenseVoice and Paraformer inference

[LauraGPT]

FunASR v1.3.14 is available on PyPI and GitHub.

Upgrade

pip install --upgrade funasr==1.3.14

Runtime fixes

• SenseVoice Triton import: fixes a Python SyntaxError that prevented the optimized DTW module from loading during auto-registration. (#3027)
• Paraformer short/silent segments: safely handles CIF outputs with no fired boundary instead of crashing. On the reported 184-file Chinese benchmark, decoding improved from 179/184 to 184/184 while micro-average CER remained 10.17%. (#3028)

Documentation

The Quick Start and model-selection path are easier to scan, benchmark evidence is linked more clearly, and the English and Chinese entry points are synchronized.

• Release notes
• Full changelog
• Model selection guide

When reporting a regression, please include the FunASR version, model ID, device, minimal command or code, and the complete traceback. A short reproducible audio sample is especially helpful when it can be shared.